Download On the Use of Actor-Network Theory for Developing Web Services

On the Use of Actor-Network Theory for Developing Web
Services Dedicated to Communities of Practice
Liliane Esnault - [email protected]
EM LYON – 23 Avenue Guy de Collogue – 69134 Ecully Cedex – France
Romain Zeiliger - [email protected]
Frederic Vermeulin - [email protected]
GATE-CNRS – 13 Chemin des Mouilles – 69132 Ecully Cedex – France
Abstract. Communities of practices are more and more recognized by companies,
individuals and groups as valuable places to share and create knowledge. Communities of
practice have to be cultivated in order to fully create the value they may bring to their
environment. They need interoperable, flexible, ubiquitous, and specific collaborative
tools to support their work. Developing such tools and enabling their use among
communities of practice requires adopting a Participatory Design approach. Actor
Network Theory (ANT) is then used to define a methodology that fosters the
participation of heterogeneous stakeholders to the design process. As a case study, we
show that ANT concepts are useful to analyse the design context of an European project
named PALETTE that aims at developing interoperable services for helping communities
of practices to better cultivate themselves
Keywords: communities of practice, participatory design, actor-network theory,
European projects
1 Introduction
Our focus is on reflecting upon a design methodology that could help providing a
community of practice (CoP) with enough suitable and usable tools so that it is able to
cultivate itself appropriately. Wenger et al. (Wenger et al. 2002) identify seven
principles for cultivating communities of practice: design for evolution, open a dialog
between inside and outside perspectives, invite different levels of participation,
develop both public and private community spaces, focus on value, combine
familiarity and excitement, create a rhythm for the community.
We are well aware that these processes of cultivating communities of practice go far
beyond the design of tools. But collaboration, communication, knowledge
management, document exchange, problem solving are activities that cannot be
accomplished without a strong support from technology enabled tools, all the more so
because communities members are scattered in different locations and even across
different organizations.
Usual communication tools like e-mail, and forums are naturally the common basis for
communication. But they are not efficient enough to really support the development of
the activities of a community of practice to the point where it can start create value for
E. Tomadaki and P. Scott (Eds.): Innovative Approaches for Learning and Knowledge Sharing,
EC-TEL 2006 Workshops Proceedings, ISSN 1613-0073, p. 298-306, 2006.
Actor-Network Theory for Developing Web Services Dedicated to Communities of Practice
299
itself and its environment. CoPs need tools that answer better to their specific needs
and usages. They have - simply - to create, reuse, store, share, exchange, publish,
represent and capitalize information. But the nature of the information, the content and
the value of it are somewhat different than what a usual database or document
management system can provide. The information that is manipulated in a community
of practice is more often informal, declarative, textual or graphic, qualitative, practiceoriented, sometimes not well formalized or even expressed. It may deal with a lot of
different subjects, the relevance and value of which are only decided by the
community. Thus a community of practice needs tools that share some common
features, among which:
• be available anywhere;
• allow flexible use, depending of the skills of the members regarding technology;
• cover a range of document management functions: creation, modification,
publication, exchange, storage, retrieval, all on a collaborative basis;
• cover a range of information representation and modelling functions providing a
mean for creating a common ground within the community;
• cover a range of knowledge management functions, related to the practice and the
identity of the community, and the learning activities within the community;
• enable communication, collaboration and cooperation in the way that is useful for
the community, both inside the community and between the community and its
environment;
• allow to understand, represent, enrich, share members' expertise.
Such tools might be different - even radically - from the usual IT that are used in
companies or for the day-to-day job. They are merely based on new technologies,
open-source or "open-source minded" (the usefulness and quality of which are
qualified by users, not by proprietary developers). They have to be interoperable,
evolutionary, flexible and truly collaborative. They are likely to appear as a set (a
"palette" of interoperable web services.
2 Designing web services for communities of practice: ActorNetwork Theory and Participatory Design
The Participatory Design approach may be considered as a process of negotiation of
usefulness to be achieved through reconciling the contrasting perspectives of various
stakeholders, including users, designers and others. We argue that using ANT1 to
1
ANT was formerly the acronym for Actor-Network Theory. It
is now used as itself, and even one of its first creators, Bruno
Latour, recognises that it has become something different, and if
it was created now, he would probably not have used the same
words, specially the word network that he feels confusing now
[Latour 1999]. We will then use ANT as a name and not as an
acronym
300
L. Esnault, R. Zeiliger, and F. Vermeulin
analyse design situations where CoPs2 are involved can help defining a suitable
Participatory Design methodology.
There are different interpretations of the usefulness of technology. As stated by Abreu
de Paula: "perception of usefulness is not statically embedded in its design, but is
dynamically and constantly created and shaped by different social groups. In this
respect, one important goal is to attempt to reconcile these often contrasting
perspectives" (Abreu de Paula, 2004). While Participatory Design does not explicitly
address the social construction of usefulness, it may be considered as framing the
social interactions that eventually lead to a recognised useful system.
The main difficulty of Participatory Design remains the organization and management
of an efficient participation – i.e. a participation that can truly influence the design
process. Each actor of the design process is an expert of her domain and this expertise
influences the design process. However actors are heterogeneous in respect to their
disciplines, preoccupations and interests: they don’t speak the same "language". For
them to interact necessitates that they construct together a "common ground". This is
achieved through participative activities that mediate participation. Examples of such
activities include brainstorming meetings, prototype demonstration, scenario
performing, role playing, design games. Participative activities are often hampered by
suspicion and even conflict.
Some of these activities may focus on creating boundary-objects (Bowker and Star,
1999; Gasson, 2006) i.e. objects "to-think-with" that facilitate mutual understanding
and trust among participants with various backgrounds. A mock-up, an intermediate
version of the final product, a use-case or a scenario are classical boundary-objects.
This concept is closely related to what Wenger says about reification: "reification …
refer to the process of giving form to our experience by producing objects (…) In so
doing we create points of focus around which the negotiation of meaning becomes
organized"(Wenger, 1998).
ANT provides a conceptual framework helping formulating and building a design
methodology that sustains efficient participation of heterogeneous. ANT portrays an
alignment that differs from the traditional system development one along crucial
dimensions: there is an open-ended array of "things" that need to be aligned including
work-routines, incentive structures, system modules and organisational roles. It follows
immediately that there can be no strict top-down control over such a collection of
things (Monteiro, 2000). Actors' heterogeneity is one of ANT main originalities. An
actor is characterized first hand by its capability to act and interact, its influence. ANT
thus clearly acknowledges that a lot of "things" - humans and non-humans - do have an
influence (McBride). The notion of participation is extended to take into account the
2
Just as for ANT, we will use the term CoP to refer to a
community of practice, following Wenger's recommendation
that "community-of-practice" should be viewed as a unit
(Wenger, 1998)
Actor-Network Theory for Developing Web Services Dedicated to Communities of Practice
301
participation/influence of non-human actors, such as artefacts and organisations. This
is obviously an interesting feature when describing a socio-technical system.
ANT concepts seem appropriate for preparing design strategies, in a Participatory
Design context, that aim at "aligning the interests of the actor-network " i.e. having all
their influences fit together. The alignment of the network is obtained through
processes of translation: translation means both a move of some actor's interests and a
translation - in the sense of change of language or representation - of those interests in
order to align them with the interests of other actors. According to Callon (Callon,
1999), the translation process includes several steps, among which: interessement and
enrolment. Interessment and enrolment focus on negotiating acceptable roles for the
human actors.
The next ANT concept is inscription, meaning that "aligned interests [are] inscribed
into durable material" (Law, 1992). A translation process supposes a medium or a
material in which it is inscribed (boundary objects, for example, may support
inscription). According to Akrich: "A large part of the work of innovators is that of
inscribing their vision of the world in the technical content of a new object" (Akrich,
1992).
Finally, ANT introduces the concept of black-boxing. Back-boxes are "sealed actornetworks" (Stalder, 1997) whose alignment has been obtained, whose aligned interests
have been inscribed in a stable association that is no longer questionable – except at a
heavy cost. In this sense, a project plan is a black-box that has been sealed after a
translation process has succeeded in aligning the interests of the project partners.
3. A case Study: the PALETTE European Project.
As an illustration of the concepts described above, we would like to present the context
of a European Project named PALETTE3. It gathers about fifty researchers from
thirteen Institutions in seven countries. The PALETTE project aims at facilitating and
augmenting individual and organisational learning in communities of practice. To
reach this aim, an interoperable and extensible set of innovative services as well as a
set of specific scenarios of use will be designed, implemented and validated in CoPs of
diverse contexts. The PALETTE services are classified into three categories:
information services, knowledge management services and mediation services.
PALETTE adopts a participative design approach, establishing a good balance between
technological and pedagogical experts. Evaluation is integrated in the same process, in
order to provide direct, frequent and detailed feedback.
It is expected that the adoption of the developed services and scenarios will result in
3
(6th Framework Programme - Priority IST-2004-2.4.10 Technology-Enhanced Learning). PALETTE stands for:
Pedagogically sustained Adaptive LEarning Through the
exploitation of Tacit and Explicit knowledge
302
L. Esnault, R. Zeiliger, and F. Vermeulin
•
•
•
•
•
•
•
the facilitation of tasks performed by learning CoPs by removing barriers
imposed by current approaches;
the exploitation of diverse mental models, knowledge resources and
competences of CoPs member through the social interaction of codified and
tacit knowledge;
the creation of new knowledge, which can lead to the evolution of the
associated learning resources;
the easy access and reuse of knowledge built by the CoPs;
the increase of active participation of individuals in CoPs;
the emergence of new CoPs, inside and outside organisations;
the increase of the overall quality of learning in CoPs.
PALETTE will provide innovative models and technical solutions with regard to the
following dimensions:
• efficient reuse and sharing of information among the CoPs' participants;
• user-friendly production and use of multimedia content to support the
expression of practices (behaviour, rules, personal theory, etc.);
• efficient and effective support of the individual and organisational learning
process, the incoming of new participants in a CoP, and the capitalization of
knowledge.
PALETTE will implement the conditions for the exploitation and development of open
source services by a large number of CoPs. Thus the PALETTE services and scenarios
will not only address the needs of identified CoPs but also describe the conditions for
their enhancement through the active participation of users in their development.
One of the first tasks of the project, which started in February 2006, was to settle a
design methodology implementing the conditions for Participatory Design. It seemed
that ANT was a good support for creating a common understanding of the
methodological context of the project.
3.1 Using ANT to implement the participatory design methodology in Palette.
McBride (McBride) suggests a 7 steps methodology where ANT is used as an
analytical tool "to identify actions which may speed the social embedding of the
technology and the successful take-up of (a) system": identify stakeholders, investigate
stakeholders, identify stakeholders' interactions, build actor-network models, identify
irreversibility (provisional stabilities), identify inhibitors and promoters, identify
actions for aligning the network (participative activities).
In PALETTE, we apply an analogous methodology to implement the Participatory
Design process:
•
the first steps consist in identifying the various stakeholders, their interests, the
inhibitors and the promoters for the enrolment of these actors in the actor-network;
Actor-Network Theory for Developing Web Services Dedicated to Communities of Practice
•
•
303
then, by attempting to "align" these actors' interests, we will build the actornetwork and an ANT-based description of the issues related to bootstrapping the
participatory approach in Palette;
finally we will propose a set of actions – mainly participative activities with
boundary objects – and select a set of inscription medium with the aim to "enrol"
the various actors and promote the social design and acceptance of the new
technologies.
3.2 Building the actor-network: identifying and enrolling the actors, aligning their
interests
There are a lot of actors gathered for the project purposes.
• CoPs, CoPs members, CoPs animators
• CoPs observers, community of CoPs observers
• Project, DoW4, project coordination, project management
• Research teams
• Work Packages, tasks groups, sub-tasks groups
• Pedagogical tools: social sciences methodologies, interviews, scenarios, data
collection methods, data representation methods…
• Methodological tools: ANT, MOT…
• Management tools: (reports, time-sheets, deliverables)
• Technical tools: from the project (existing and potential) and existing outside the
project
• Technical tools designers and developers (called "Ts" in the project)
• Pedagogical tools designers (called "Ps" in the project)
• Methodological tools providers (Ts + Ps)
Most of them already existed before the project and will continue their life after the
project: researchers, institutions, currently existing tools, some CoPs, etc.. Some of
these actors had already build relationship between themselves, some other not. Some
actors will exist only due to the project: the newly developed tools, the Work Packages,
the deliverables, for example. The PALETTE actor-network is a dynamic entity which
is made of all the heterogeneous actors (meaning human and non human, but also of
different granularity5) and of all the links that tie dynamically these actors for the
purposes of the project (and also for other possible reasons).
The situation of an actor within an actor network is not fully defined by the existence
of the actor. Some links have to be knitted with other actors to materialize the presence
of the actor in the network, through enrolment. Enrolling an actor within an actornetwork means that there are some agreed common interests between this specific actor
and the actor-network at some moment. Building the partnership between institutions
4
Description of Work, the reference document for the project
A workpackage, or an institution, or a project is made of a lot
of persons and other elements; thus a person and a group of
person are actors of a different level of granularity
5
304
L. Esnault, R. Zeiliger, and F. Vermeulin
(in fact groups within institutions) to submit a proposal to the European Call for
Projects was a first kind of enrolment
Enrolling actors in an actor-network requires going through some participative
activities where actors can discover and share their common interests. The CoPs are not
members of the project, but it is really important that they become actors of the project.
Thus, They have to be enrolled, by identifying some common interest between CoPs,
and/or CoPs' members, and other actors of the PALETTE actor-network. The
Participative Interview process that is used to gather data about the CoPs is the main
step toward enrolling them.
Currently existing collaborative tools (like Lotus Notes or e-Rooms, or Moodle, etc.)
are not partners of the project as well. But they are used by a lot of people and by CoPs
outside the project. They have to be taken into account in the project, from a technical
point of view - which is a matter of interoperability and standards - and from a user
interface point of view as well. This is done through the Tool Inventory/Categorization
process, which is the main participative activity through which tools are enrolled in the
PALETTE actor-network. For "inside" tools (those developed by partners), the
categorization is not the only enrolment process; another enrolment process is that they
are used within the project (for example, a document management software is used to
collaboratively publish project documents)
3.3 Inscribing aligned interest in scenarios of use
Translation and inscription are dual processes. In PALETTE, a successive number of
translations are undertaken from CoPs to CoPs observers, then to interviews
transcriptions, then to data condensation; the data are finally inscribed in data
representation supports available as boundary objects for other actors. Different media
are used for inscribing, like documents, story telling, vignettes, and MOT schemas
(Paquette et al., 2006). Another example of the translation-inscription process is the
activity aiming at clarifying the notion of scenario: what is a scenario, its content, its
form, etc., according to the different PALETTE actors. Several participative activities
are designed to make explicit the representations/interests of the actors and
progressively "inscribe" a definition and typical contents/forms of scenarios useful for
all the actors, according to [Iacucci & Kuutti].
3.4 PALETTE incremental project life cycle.
The organisation (structure and stages) of a project life-cycle is a key factor for the
success of a Participative Design approach. Let say it in ANT terms: the building of the
actor-network, its evolution throughout the project, the nature of the translationinscription processes, as well as the nature and number of boundary objects depend on
the type of project life-cycle. The project methodology used in PALETTE is based on
an agile perspective (Schwaber, 2004, Highsmith, 2004): go for a "first design round",
Actor-Network Theory for Developing Web Services Dedicated to Communities of Practice
305
with some "sample" CoPs, a few data from interviews, a few tools, and try to build
from this the first scenarios. This would allow us:
• to validate the methodology
• to validate the feasibility of the whole process;
• to understand better what the different steps are (especially the data representing
and the scenario building);
• to explicit the processes of enrolment, translation and inscription and see if it fits
really well and if everybody agrees with;
• to go further into inscribing communities practices and web services into scenarios
of use.
From these scenarios, some adjustments (including possibly incremental or full new
developments) could be done in the partners' tools to better suit CoPs users' practical
situations. Then we will be able to re-loop the loop with other CoPs and other tools (to
keep it simple, though there are all other actors involved). Step by step we will build, in
a constructive perspective, our scenarios and use-cases.
The multi-rounds project life cycle allows going on rather quickly in the validation of
the whole project system and enables actors working more collaboratively from the
beginning.
4 Conclusion and further research
PALETTE has just started in February 2006. We are still in the process of looping the
first design round. What we were able to agree upon so far is that descriptions of
design situations based on ANT concepts have helped launching an efficient
Participatory Design methodology. The inscriptions as MOT schemas, for example,
were agreed a "good" boundary objects by both the Pedagogical and Technical
partners. A lot has still to be achieved before the end of the project in January 2009.
Nevertheless, we think that PALETTE is a good example of a complex socio-technical
project, and that this experience of using ANT could benefit other kinds of complex
socio-technical projects.
References
1.
2.
3.
4.
Abreu de Paula, R. 2004, The Construction of Usefulness: How Users and Context Create
Meaning with a Social Networking System – Dissertation available at
http://www.ics.uci.edu/~depaula/publications/dissertation-depaula-2004.pdf.
Akrich M., 1992, The des-cription of technical objects, in Bijker, W.E., and Law, J.(editors)
Shaping Technology / Building Society , MIT Press
Bowker; G.C., and Star, S.L., 1999, Sorting Things Out, Classification and its
consequences, MIT press, Cambridge, MA
Callon, M. (1999). Some Elements of a Sociology of Translation: Domestication of the
Scallops and the Fishermen of Saint Brieuc Bay. In M. Biagioli (Ed.) The Sciencer Studies
Reader. New York and London, Routledge: 67-83.
306
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
L. Esnault, R. Zeiliger, and F. Vermeulin
Gasson, S., A genealogical study of boundary-spanning IS design, European Journal of
Information Systems, Vol 15, N° 1, February 2006, pp 26-41
Highsmith, Jim, The Agile Revolution, Addison Wesley, 2004
Iacucci, G., and Kuutti, K., 2002, Everyday Life as a Stage in Creating and Performing
Scenarios for Wireless Devices, in Personal and Ubiquitous Computing, Springer-Verlag
Latour, B., 1999, On Recalling ANT, in Actor network Theory and After, John Law and
John Hassard editors; Blackwell Publishing, Oxford, UK.
Latour, B. 1992, The Sociology of a Few Mundane Artifacts, in Bijker, W.E., and Law, J.
(editors) Shaping Technology / Building Society , MIT Press
Latour, B., 1996, Aramis or The Love of Technology, Harvard University Press,
Cambridge, MA
Law, J., 1992, Notes on the Theory of the Actor-Network: ordering, strategy, and
heterogeneity, Systems Practise, 5(4), pp379-393
Maner, Walter, http://csweb.cs.bgsu.edu/maner/domains/RAD.htm
McBride http://www.cse.dmu.ac.uk/~nkm/WTCPAP.html
Cooney, E.M., Millery, M., Larson, H., and Choriki, D., User Friendly: Dialectic
Development of Technology-Enhanced Learning Environments, Workshop of the 2000
Participatory Design Conference
Monteiro, Eric, Actor-network theory. In: C. Ciborra (ed.), From Control to Drift. The
Dynamics of Corporate Information Infrastructure, Oxford Univ. Press, 2000, pp. 71 – 83
Paquette, G., Léonard, M., Lundgren-Cayrol, K., Mihaila, S., & Gareau, D. (2006).
Learning Design based on Graphical Knowledge-Modelling. Educational Technology &
Society, 9 (1), 97-112.
Stalder, Felix, 1997 http://felix.openflows.org/html/Network_Theory.html
Tilson, David and Lyytinen, Kalle, Making Broadband Wireless Services: An ActorNetwork Study of the US Wireless Industry Standard Adoption, Sprouts: Working Papers
on Information Environments, Systems and Organizations, Volume 5 Issue 3, 2005, pp 137154
Wenger, Etienne, Communities of Practice: Learning, Meaning and Identity, Cambridge
University Press, Cambridge, UK, 1998
Wenger, E;, McDermott, R, Snyder, W.M., Cultivating Communities of Practice, Harvard
Business School Press, 2002